UAV-based deep learning for biodiversity monitoring: advances, applications, and future directions

Biodiversity forms the foundation of Earth’s ecosystem health and human well-being. However, it is currently under unprecedented threat. Traditional biodiversity monitoring methods are often constrained by limitations in time, space, and manpower, making it difficult to meet the demands for large-scale, high-frequency, and high-precision monitoring. This review focuses on an emerging technological integration - the combination of active electro-optical payload systems on drones and deep learning algorithms and explores its potential applications in biodiversity monitoring. We provide a detailed analysis of multi-source remote sensing data acquisition strategies, the application of deep learning algorithms for species identification and ecological feature extraction, active computing and flight path planning mechanisms, and nonlinear ecological feature analysis methods. Through case studies, we demonstrate the practical application of this technological integration in monitoring flagship species in national parks. The results indicate that this innovative approach significantly enhances the efficiency, accuracy, and coverage of biodiversity monitoring, providing robust support for ecosystem management and conservation decisions. Despite challenges such as hardware limitations, algorithm optimization, and ecological interpretation, the continued advancement of technology positions intelligent drone-based monitoring systems to play an increasingly important role in biodiversity research and conservation, providing novel solutions to address the global biodiversity crisis. • Provides a comprehensive review of deep learning applications in biodiversity monitoring. • Summarizes advances in UAV, remote sensing, and acoustic monitoring technologies. • Discusses emerging AI paradigms including Transformers, GNNs, GANs, and self-supervised learning. • Identifies hardware, algorithmic, and ecological limitations in current monitoring systems. • Proposes future directions for interpretable, integrated, and sustainable AI-driven ecology.